metadata
base_model: microsoft/trocr-base-handwritten
tags:
- generated_from_trainer
model-index:
- name: license_plate_recognizer
results: []
license_plate_recognizer
This model is a fine-tuned version of microsoft/trocr-base-handwritten on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.0097
- Cer: 0.0036
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 23
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 0.1485 | 1.0 | 397 | 0.0528 | 0.0182 |
| 0.0843 | 2.0 | 794 | 0.0371 | 0.0089 |
| 0.0552 | 3.0 | 1191 | 0.0417 | 0.0129 |
| 0.0812 | 4.0 | 1588 | 0.0386 | 0.0115 |
| 0.0315 | 5.0 | 1985 | 0.0198 | 0.0053 |
| 0.0178 | 6.0 | 2382 | 0.0263 | 0.0084 |
| 0.0341 | 7.0 | 2779 | 0.0179 | 0.0067 |
| 0.0143 | 8.0 | 3176 | 0.0149 | 0.0080 |
| 0.0047 | 9.0 | 3573 | 0.0055 | 0.0027 |
| 0.0163 | 10.0 | 3970 | 0.0062 | 0.0022 |
| 0.0045 | 11.0 | 4367 | 0.0049 | 0.0027 |
| 0.0115 | 12.0 | 4764 | 0.0077 | 0.0053 |
| 0.0014 | 13.0 | 5161 | 0.0031 | 0.0022 |
| 0.0081 | 14.0 | 5558 | 0.0052 | 0.0031 |
| 0.0001 | 15.0 | 5955 | 0.0056 | 0.0035 |
| 0.0005 | 16.0 | 6352 | 0.0057 | 0.0027 |
| 0.0009 | 17.0 | 6749 | 0.0053 | 0.0022 |
| 0.0003 | 18.0 | 7146 | 0.0067 | 0.0027 |
| 0.0001 | 19.0 | 7543 | 0.0044 | 0.0018 |
| 0.0001 | 20.0 | 7940 | 0.0052 | 0.0018 |
| 0.0 | 21.0 | 8337 | 0.0050 | 0.0018 |
| 0.0 | 22.0 | 8734 | 0.0051 | 0.0018 |
| 0.0 | 23.0 | 9131 | 0.0051 | 0.0018 |
Framework versions
- Transformers 4.42.3
- Pytorch 2.1.2
- Datasets 2.20.0
- Tokenizers 0.19.1